Our study shows that it is possible to completely fill the void left by an irreparable rotator cuff tear with a GRAST graft. The braiding of the GRAST makes it possible to reproduce the morphology of the torn tendon in all three dimensions of the space, all the better when the GRAST is hydrated (as is the case once it is implanted, especially under arthroscopy). The GRAST graft was able to connect the greater tuberosity of the humerus to the spine of the scapula in all cases. Finally, the passive mechanical behavior of the GRAST does not interfere with the elevation of the arm.
Our technique offers an intermediate solution between superior capsular reconstruction, which has become the primary treatment in many centers in the management of irreparable posterosuperior tears [14], along with partial repair with medialization [5] when the patient is not eligible for a reverse shoulder arthroplasty or tendon transfer. It simplifies medial fixation and restores the musculo-tendinous chain where current grafting techniques are limited with completely filling a tendon defect. In fact, extending the graft to the supraspinatus fossa also offers the hope of incorporation into the body of the muscle.
Centering the head of the humerus in the glenoid is an essential prerequisite for the proper function of the shoulder, especially with forward elevation. Indeed, this allows the center of rotation of the humeral head to be maintained in an optimal position so that contraction of the deltoid causes rotation of the humeral head rather than proximal migration. In addition, it induces a tension effect on the deltoid fibres, which potentiates its strength when the elevation movement is initiated. The natural tendency of the humeral head is to migrate proximally under the action of the powerful deltoid. In the event of a massive tear in the cuff, this results in a reduction of the subacromial space. Since glenohumeral congruence is poor, only soft tissues (ligaments, tendons) can oppose this migration. As in many other anatomical sites, there are two complementary solutions to prevent migration of a bone piece in a given direction: “traction” in the opposite direction and “abutment”. Traction is usually based on a ligament and/or tendon whose force vector is in the opposite direction to the movement to be prevented. This is similar to the cable of an anchor that prevents the boat from drifting under the action of the current. The stop, on the other hand, consists of placing an obstacle that will oppose the migration of the bone piece. In the shoulder, the only stop between the humeral head and the acromion is the supraspinatus tendon. It has therefore been suggested that proximal migration of the humerus could be prevented by sliding a subacromial spacer implant. This is the case for example with subacromial balloons [16, 17]. However, when the arm is raised, the greater tuberosity will engage under the acromion and the subacromial space will become almost obsolete. If this space is already occupied by a subacromial spacer, the course of the greater tuberosity will be blocked because it will come into abutment with the spacer. The same structure that keeps the humerus at a distance from the acromion when the arm is in the lowered position must therefore be able to be gradually moved out of the way when the arm is raised, without blocking the course of the greater tuberosity. This is exactly what the supraspinatus tendon does. Our study confirms that during passive elevation of the arm, the GRAST graft gradually escapes from the subacromial space into the supraspinatus fossa. (Fig. 5).
However, this study is a cadaveric study and therefore has certain limitations.
As with any anatomical study, the question of in vivo transposition arises. Indeed, one legitimate question is the clinical tolerance of such a transplant as well as the biological fate of the transplant. It is obviously too early to report the results of a clinical series, but a first look at early results can be seen with the case of a 73 years old patient whom we treated for a painful shoulder with irreparable cuff tear. We inserted a GRAST graft using arthroscopy. At 1 year’s follow-up, the graft appeared well incorporated with a reconstructed cuff appearance. Clinically, only intermittent residual pain persists, and it is interesting to note that the GRAST does not interfere with the elevation of the arm, which is almost complete (Fig. 6). This is in line with our biomechanical hypothesis on the behavior of GRAST.
After a rotator cuff tear, inactive muscles tend to atrophy and degenerate into fat, visible on Computerized Tomography (CT) and Magnetic Resonance Imaging (MRI) [13]. It would be interesting to evaluate whether restoring continuity between the posterosuperior cuff and the greater tuberosity reactivates the function of an atrophied muscle. Our only clinical case does not allow us to conclude such a recovery, but we have started a clinical study by including patients operated on using the Braided GRAST technique under arthroscopy, by comparing muscle trophicity with imaging of the supraspinatus before and at a distance from surgery, and by evaluating the contractility of the supraspinatus muscle whose continuity with the greater tuberosity has been restored in dynamic ultrasound.
A second limitation is the need to harvest from another anatomical site. As with any graft harvesting, this adds complexity to the technique and increases the risk of specific complications: infection, donor site morbidity, failure of the harvesting, residual pain. This limitation is important as more and more surgeons are over-specialized in shoulder surgery and are therefore unfamiliar with the removal of harmstring tendons. However, the removal of these tendons is commonly performed for cruciate ligament surgery. For surgeons who are used to this procedure, it takes only a few minutes and the scar is short. The rate of infection and specific complications is exceptionally low [7]. The main complication is damage to the infrapatellar branch of the saphenous nerve, generating hypoesthesia distal to the scar. Although frequent, estimated at between 30 and 50% of patients, this complication is extremely well tolerated [9]. Furthermore, the GRA and ST tendons offer a tissue volume that is much greater than that which the upper limb can provide, which explains their varied use for ligament surgery of the upper limb [9, 18]. Finally, these tendons regrow between 60 and 80% of cases in the months following their removal [19].